triton-lang triton issues

triton-lang / triton

Development repository for the Triton language and compiler

https://triton-lang.org/

MIT License

13.54k stars 1.67k forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

Support computation pipelining after SWP refactoring

#5185 manman-ren opened 1 week ago
1
[BACKEND] Add missing precondition in optimize acc init

#5184 ThomasRaoux closed 1 week ago
0
[LAYOUTS] Unify the implementation of getShapePerCTATile

#5183 lezcano closed 1 week ago
0
The lastest triton cannot run tutorials

#5182 zhaosiying12138 closed 1 week ago
1
[BACKEND] Fix ProgramPoint passing in AxisInfoAnalysis

#5181 aakhundov closed 1 week ago
3
Update to llvm/llvm-project@bd9145c8c213

#5180 antiagainst closed 1 week ago
0
Update to llvm/llvm-project@bd9145c8c213

#5179 antiagainst closed 1 week ago
0
[Persistent][Pipelining] Fused 08-grouped-gemm 2 inner loops for better pipelining, got bad performance

#5178 CharlieFRuan opened 1 week ago
1
[AMD] Enable mixed precision matmul test

#5177 makslevental closed 1 week ago
0
[PIPELINER] Cleanup of LoopScheduling.cpp, introduction of AssignLatencies

#5176 pawelszczerbuk opened 1 week ago
1
[AMD][Pipeliner] Reland "Improve clustering and add prefetch"

#5175 sjw36 closed 1 week ago
0
[Persistent] Performance on 09-persistent-matmul on A100 worse than non-persistent

#5174 CharlieFRuan opened 1 week ago
1
[Triton] Default diagnostic handler only filters for errors

#5173 Mogball closed 1 week ago
0
[Persistent][Pipeline] Moving `tile_id += NUM_SMS` to epilogue fails to pipeline in 09-persistent-matmul

#5172 CharlieFRuan opened 1 week ago
4
[CI] Disable `MLIR_ENABLE_REMARK`

#5171 Jokeren closed 1 week ago
0
[LAYOUTS] Implement IR support for LinearLayouts

#5170 lezcano closed 1 week ago
6
[INTERPRETER] Fix argument passing for internal parameters in function declarations

#5169 Jokeren closed 1 week ago
0
[AMD] NFC: Drop duplicated moveUpTranspose

#5168 antiagainst closed 1 week ago
0
[BACKEND] Cleanup redundant broadcast combine pattern

#5167 peterbell10 closed 1 week ago
0
[BACKEND] Add folder for `addptr(ptr, 0) -> ptr`

#5166 peterbell10 closed 1 week ago
0
[BACKEND] Update LLVM version to https://github.com/llvm/llvm-project/commit/fb4f426c81d7e87dbb30df7abeba15ffc2f9f41a

#5165 vwbaker closed 1 week ago
1
Setting the environment variable TRITON_INTERPRET causes the kernel function to not be able to receive reserved keyword arguments.

#5164 0Addicted0 closed 1 week ago
0
[AMD] Add instruction schedule loop boundary guard hints

#5163 ravil-mobile opened 2 weeks ago
3
[Tutorial] Remove incorrect caching from softmax tutorial

#5162 Mogball closed 2 days ago
0
[TritonGPU] Fix incorrect mask operand used in for loop pipeliner

#5161 Mogball closed 1 week ago
0
Restore CentOS 7 build and backfill for release/3.2.x

#5160 bertmaher closed 2 weeks ago
0
Add Sageattention Codes as a tutorial

#5159 jt-zhang closed 2 weeks ago
2
Restore the CentOS 7 build

#5158 bertmaher closed 1 week ago
4
Revert "[AMD][Pipeliner] Improve clustering and add prefetch (#4881)"

#5157 antiagainst closed 2 weeks ago
0
[AMD] Reland "pipeliner clustering and prefetch"

#5156 antiagainst closed 1 week ago
0
[Triton] Generate local MLIR reproducers when possible

#5155 Mogball closed 2 weeks ago
9
[BACKEND][DRAFT] Use linear layout for loading mmav2 dot operand tensors from shared memory

#5154 Jokeren opened 2 weeks ago
0
[AMD] Fix slow compilation issue due to inline print calls

#5153 binarman closed 6 days ago
0
[Triton] Remove upstream bug workaround (NFC)

#5152 Mogball closed 2 weeks ago
0
[TEST] Make mixed matmul test deterministic

#5151 ThomasRaoux closed 2 weeks ago
0
Create an aggregate `check-triton-unit` target

#5150 Mogball closed 2 days ago
0
Fix `gtest_discover_tests` timeout argument

#5149 Mogball closed 2 weeks ago
0
[AMD] inThreadTranspose: Transpose between global load and local store for non-TN layouts: part 1 of 4

#5148 jtang10 opened 2 weeks ago
0
[IR] Add typing for tensor descriptor types

#5147 peterbell10 closed 2 weeks ago
0
Load backend dialects in `IRSource` to make sure `parse_mlir_module` works for third_party backends

#5146 anmyachev closed 1 week ago
2
Use pytest' `tmp_path` in `test_irsource.py`

#5145 anmyachev closed 2 weeks ago
0
[AMD] Refactoring Instruction Scheduling

#5144 ravil-mobile closed 1 week ago
2
feat: Dev Container for consistent dev setup

#5143 maryamtahhan opened 2 weeks ago
3
RuntimeError: Cannot call @triton.jit'd outside of the scope of a kernel

#5142 yang9936 closed 1 week ago
10
有人遇到过yolov8n.pt模型转torchscripts和onnx，在triton server或Deepytorch Inference上推理，精度下降的问题吗？

#5141 JackonLiu closed 2 weeks ago
1
[PROTON] Fix proton's support for multiple profiling sessions

#5140 Jokeren closed 2 weeks ago
0
[AMD] Use warp shuffle for MFMA to Dot operand layout conversion (FP8)

#5139 ilia-cher closed 1 week ago
4
triton GEMM with size < 16

#5138 March-H opened 2 weeks ago
0
[AMD] Get rid of flat load/store instructions

#5137 joviliast closed 2 weeks ago
1
Warp memory alignment error when manually launching compiled PTX

#5136 noctrog closed 2 weeks ago
2

Previous Next